Hello Russell,
I am testing HFSC in a virtual lab at home. I have Ostinato packet generator sending continuous flows to a host, I have a GNU+Linux machine in between, where I have configured HFSC. Everything is shaped perfectly by HFSC, but there is a problem with delay. I would like to know what I am doing wrong because traffic on the class where I configured rt has more than double delay than any other traffic. In order to check delay I use qperf and ping on the top of the Ostinato flows.
I would be very happy if you could let me know whether you see something wrong in my configuration.
me@gnulinux# tc qdisc show dev eth0 qdisc hfsc 1: root refcnt 2 default 15 qdisc sfq fa: parent 1:a limit 127p quantum 1514b depth 127 divisor 1024 perturb 10sec qdisc sfq f15: parent 1:15 limit 127p quantum 1514b depth 127 divisor 1024 perturb 10sec qdisc sfq f1: parent 1:1 limit 127p quantum 1514b depth 127 divisor 1024 perturb 10sec qdisc sfq f14: parent 1:14 limit 127p quantum 1514b depth 127 divisor 1024 perturb 10sec me@gnulinux# tc class show dev eth0 class hfsc 1: root class hfsc 1:1 parent 1: sc m1 0bit d 0us m2 1500Kbit ul m1 0bit d 0us m2 1500Kbit class hfsc 1:a parent 1:1 leaf fa: rt m1 250Kbit d 8.0ms m2 100Kbit class hfsc 1:15 parent 1:1 leaf f15: ls m1 0bit d 0us m2 150Kbit class hfsc 1:14 parent 1:1 leaf f14: ls m1 0bit d 0us m2 150Kbit ul m1 0bit d 0us m2 300Kbit class sfq fa:2e2 parent fa: class sfq f15:232 parent f15: class sfq f15:2c8 parent f15: class sfq f14:320 parent f14:
By the way, I am not sure if it is ok to ask here this kind of questions. Please let me know if you think I should address my HFSC support questions somewhere else, this is the only place I found...
Thank you very much!
I think I can ask it in an easier way.
With the configuration below I get far more delay on 192.168.10.0/24 traffic than on 192.168.20.0/24 traffic.
Am I doing something wrong?
If you think it is better to ask for support somewhere else and you know where, I'd be grateful if you let me know.
Thank you.
I have also tested the examples at http://linux-tc-notes.sourceforge.net/tc/doc/sch_hfsc.txt in two different labs (both virtuals, I must say) and have not managed to see the expected latency results.